Back to EveryPatent.com
| United States Patent |
6,046,977
|
|
Bartkowiak
|
April 4, 2000
|
Method and apparatus for detecting signal misalignment
Abstract
A signal detector includes a transform unit and a signal alignment unit.
The transform unit is adapted to receive a plurality of frames of time
domain input samples and generate a plurality of energy values based on
the frames of time domain input samples. Each energy value is associated
with one of the plurality of frames. The signal alignment unit is adapted
to identify at least a first partial energy frame and a second partial
energy frame from the plurality of frames. The first partial energy frame
is associated with a first frequency and has a first energy value. The
second partial energy frame is associated with a second frequency and has
a second energy value. The signal alignment unit determines a misalignment
between the first and second partial energy frames based on the first and
second energy values. A method for detecting a signal is provided. The
method includes generating a plurality of energy values based on the
frames of time domain input samples. Each energy value is associated with
one of the plurality of frames. At least a first partial energy frame and
a second partial energy frame are identified from the plurality of frames.
The first partial energy frame is associated with a first frequency and
has a first energy value. The second partial energy frame is associated
with a second frequency and has a second energy value. A misalignment
between the first and second partial energy frames is determined based on
the first and second energy values.
| Inventors:
|
Bartkowiak; John G. (Austin, TX)
|
| Assignee:
|
Advanced MicroDevices, Inc. (Austin, TX)
|
| Appl. No.:
|
159028 |
| Filed:
|
September 23, 1998 |
| Current U.S. Class: |
370/210; 370/503; 370/526; 379/386 |
| Intern'l Class: |
H04M 003/00 |
| Field of Search: |
379/386,283,382,383
370/345,210,503,509,525,526
455/426
|
References Cited
U.S. Patent Documents
| 5588053 | Dec., 1996 | Xie | 379/386.
|
| 5644634 | Jul., 1997 | Xie et al. | 379/386.
|
| 5666357 | Sep., 1997 | Jangi | 370/345.
|
| 5699421 | Dec., 1997 | Nirshberg et al. | 379/386.
|
| 5722054 | Feb., 1998 | Mizutani et al. | 455/88.
|
| 5737371 | Apr., 1998 | Jaquette | 375/357.
|
| 5809133 | Sep., 1998 | Bartkowiak et al. | 379/386.
|
| 5835574 | Nov., 1998 | Lam | 379/90.
|
| 5937059 | Aug., 1999 | Kim et al. | 379/386.
|
| 5949874 | Sep., 1999 | Mark | 379/283.
|
| 5978676 | Nov., 1999 | Guridi et al. | 455/426.
|
Primary Examiner: Nguyen; Chau
Assistant Examiner: Boortalary; Forood
Attorney, Agent or Firm: William, Morgan & Amerson, P.C.
Claims
What is claimed:
1. A signal detector, comprising:
a transform unit adapted to receive a plurality of frames of time domain
input samples and generate a plurality of energy values based on the
frames of time domain input samples, each energy value being associated
with one of the plurality of frames; and
a signal alignment unit adapted to identify at least a first partial energy
frame and a second partial energy frame from the plurality of frames, the
first partial energy frame being associated with a first frequency and
having a first energy value, and the second partial energy frame being
associated with a second frequency and having a second energy value,
wherein the signal alignment unit determines a misalignment between the
first and second partial energy frames based on the first and second
energy values.
2. The signal detector of claim 1, further comprising:
a receiver adapted to sample an input signal and generate the plurality of
frames of time domain input signal samples.
3. The signal detector of claim 1, wherein the signal alignment unit is
further adapted to determine a first partial frame sample count for the
first partial energy frame and a second partial frame sample count for the
second partial energy frame, the misalignment being based on the
difference between the first and second partial frame sample counts.
4. The signal detector of claim 1, wherein the misalignment comprises a
start time misalignment.
5. The signal detector of claim 1, wherein the misalignment comprises a
stop time misalignment.
6. The signal detector of claim 3, wherein a subset of the energy values is
associated with the first frequency, and the signal alignment unit is
adapted to identify at least one full energy frame from the plurality of
frames based on the subset, the first partial frame sample count being
based on the first energy value, the number of time domain input samples
in the fall energy frame, and the energy value of the full energy frame.
7. The signal detector of claim 6, wherein the first partial frame sample
count is based on the square of the number of time domain input samples in
the full energy frame.
8. The signal detector of claim 3, wherein a subset of the energy values is
associated with the second frequency, and the signal alignment unit is
adapted to identify at least one full energy frame from the plurality of
frames based on the subset, the second partial frame sample count being
based on the second energy value, the number of time domain input samples
in the full energy frame, and the energy value of the full energy frame.
9. The signal detector of claim 8, wherein the second partial frame sample
count is based on the square of the number of time domain input samples in
the full energy frame.
10. The signal detector of claim 3, wherein the first partial energy frame
and the second partial energy frame are separated by a time interval, and
the signal alignment unit is adapted to determine the misalignment between
the first and second partial energy frames based on the first and second
energy values and the time interval.
11. A dual tone signal detector, comprising:
a transform unit adapted to receive a plurality of frames of time domain
input samples and generate a plurality of energy values based on the
frames of time domain input samples, a first plurality of the energy
values being associated with a first tone of the dual tone signal and a
second plurality of the energy values being associated with a second tone
of the dual tone signal; and
a signal alignment unit adapted to determine at least one of a first start
time and a first stop time for the first tone based on the first plurality
of energy values, at least one of a second start time and a second stop
time for the second tone based on the second plurality of energy values,
and a misalignment between at least one of the first and second start
times and the first and second stop times.
12. A signal detector, comprising:
a processor adapted to receive a plurality of frames of time domain input
samples; and
a memory device coupled to the processor and adapted to store program
instructions, that when executed by the processor:
generate a plurality of energy values based on the frames of time domain
input samples, each energy value being associated with one of the
plurality of frames;
identify at least a first partial energy frame and a second partial energy
frame from the plurality of frames, the first partial energy frame being
associated with a first frequency and having a first energy value, and the
second partial energy frame being associated with a second frequency and
having a second energy value; and
determine a misalignment between the first and second partial energy frames
based on the first and second energy values.
13. The signal detector of claim 12, further comprising:
a receiver coupled to the processor and adapted to sample an input signal
and generate the plurality of frames of time domain input signal samples.
14. The signal detector of claim 12, wherein the program instructions, when
executed by the processor,
determine a first partial frame sample count for the first partial energy
frame and a second partial frame sample count for the second partial
energy frame, the misalignment being based on the difference between the
first and second partial frame sample counts.
15. The signal detector of claim 12, wherein the program instructions that
determine the misalignment, when executed by the processor, determine a
start time misalignment.
16. The signal detector of claim 12, wherein the program instructions that
determine the misalignment, when executed by the processor, determine a
stop time misalignment.
17. The signal detector of claim 14, wherein a subset of the energy values
is associated with the first frequency, and the program instructions that
determine the first partial frame sample count, when executed by the
processor:
identify at least one full energy frame from the plurality of frames based
on the subset; and
determine the first partial frame sample count based on the first energy
value, the number of time domain input samples in the full energy frame,
and the energy value of the full energy frame.
18. The signal detector of claim 14, wherein a subset of the energy values
is associated with the second frequency, and the program instructions that
determine the second partial frame sample count, when executed by the
processor:
identify at least one full energy frame from the plurality of frames based
on the subset; and
determine the second partial frame sample count based on the second energy
value, the number of time domain input samples in the full energy frame,
and the energy value of the full energy frame.
19. The signal detector of claim 14, wherein the first partial energy frame
and the second partial energy frame are separated by a time interval, and
the program instructions that determine the misalignment, when executed by
the processor, determine the misalignment between the first and second
partial energy frames based on the first and second energy values and the
time interval.
20. A method for detecting a signal, comprising:
generating a plurality of energy values based on the frames of time domain
input samples, each energy value being associated with one of the
plurality of frames;
identifying at least a first partial energy frame and a second partial
energy frame from the plurality of frames, the first partial energy frame
being associated with a first frequency and having a first energy value,
and the second partial energy frame being associated with a second
frequency and having a second energy value; and
determining a misalignment between the first and second partial energy
frames based on the first and second energy values.
21. The method of claim 20, further comprising:
receiving an input signal; and
sampling the input signal to generate the plurality of frames of time
domain input signal samples.
22. The method of claim 20, wherein determining the misalignment includes:
determining a first partial frame sample count for the first partial energy
frame and a second partial frame sample count for the second partial
energy frame, the misalignment being based on the difference between the
first and second partial frame sample counts.
23. The method of claim 20, wherein determining the misalignment comprises
determining a start time misalignment.
24. The method of claim 20, wherein determining the misalignment comprises
determining a stop time misalignment.
25. The method of claim 22, wherein a subset of the energy value s is
associated with the first frequency, and determining the first partial
frame sample count includes:
identifying at least one full energy frame from the plurality of frames
based on the subset; and
determining the first partial frame sample count based on the first energy
value, the number of time domain input samples in the full energy frame,
and the energy value of the fill energy frame.
26. The method of claim 22, wherein a subset of the energy values is
associated with the second frequency, and determining the second partial
frame sample count includes:
identifying at least one full energy frame from the plurality of frames
based on the subset; and
determining the second partial frame sample count based on the second
energy value, the number of time domain input samples in the full energy
frame, and the energy value of the full energy frame.
27. The signal detector of claim 22, wherein the first partial energy frame
and the second partial energy frame are separated by a time interval, and
determining the misalignment includes determining the misalignment between
the first and second partial energy frames based on the first and second
energy values and the time interval.
28. A method for detecting a dual tone signal, comprising:
receiving a plurality of frames of time domain input samples;
generating a plurality of energy values based on the frames of time domain
input samples, a first plurality of the energy values being associated
with a first tone of the dual tone signal and a second plurality of the
energy values being associated with a second tone of the dual tone signal;
determining at least one of a first start time and a first stop time for
the first tone based on the first plurality of energy values;
determining at least one of a second start time and a second stop time for
the second tone based on the second plurality of energy values; and
determining a misalignment between at least one of the first and second
start times and the first and second stop times.
29. A signal detector, comprising:
means for receiving a plurality of frames of time domain input samples;
means for generating a plurality of energy values based on the frames of
time domain input samples, each energy value being associated with one of
the plurality of frames;
means for identifying at least a first partial energy frame and a second
partial energy frame from the plurality of frames, the first partial
energy frame being associated with a first frequency and having a first
energy value, and the second partial energy frame being associated with a
second frequency and having a second energy value; and
means for determining a misalignment between the first and second partial
energy frames based on the first and second energy values.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to signal detectors, and, more
particularly, to a method and apparatus for detecting a
frame-misalignment.
2. Description of the Related Art
Dual tone multifrequency (DTMF) coding is commonly used in telephone
systems to transmit information such as a phone number over the connection
between a user and a telephone central office. DTMF coding is also
becoming popular in interactive control applications, such as telephone
banking, electronic mail systems, and answering machines, wherein the user
can select options from a menu by sending DTMF signals from a telephone.
In applications where a user is dialing a phone number or providing input
to an interactive control system, a DTMF detector must discriminate
between a valid DTMF signal and speech or noise.
One type of DTMF detection error occurs when the detector erroneously
accepts signal imitations, such as those produced by speech or music, as
valid signals. This type of error is referred to as talk-off. Another type
of detection error occurs when the signal imitation interferes with a
valid DTMF signal and prevents it from being detected. This type of error
is referred to as talk-down. DTMF detectors have been developed to improve
the immunity of the detector to talk-down and talk-off effects.
Methods for detecting DTMF signals are described in U.S. Pat. Nos.
5,588,053 and 5,644,634 issued to Xie et al. (XIE), the specifications of
which are herein incorporated by reference in their entireties. In the Xie
methods, the analog signal present on the phone line is sampled and
transformed to the frequency domain to identify and extract valid DTMF
signals. The frequency domain information is analyzed to identify valid
DTMF signals and/or reject erroneous signals resulting from speech or
noise.
Current telephone systems typically offer services that send a signal to a
subscriber (i.e., telephone user) to provide identifying information
regarding an incoming call. This service is commonly referred to as Caller
ID. The communication of the identifying information occurs only if the
user is on-hook (i.e., not using the line). Another common service,
commonly referred to as Call Waiting, informs a subscriber currently using
the line (i.e., off-hook) that another call is being received.
A new feature has been developed to inform a user currently using the line
(i.e. off-hook) that another call is being received from a third party,
and to provide identifying information (i.e., Caller ID information)
regarding the third party. The user can then decide whether to take or not
take the call based on the identity of the third party. To signal the
user's telephone system of the incoming third party call, the central
office sends a customer premises equipment alerting signal (CAS) on the
user's line. The party to which the user is currently communicating does
not receive the CAS signal, but it is received by the user's telephone
system. The CAS signal is a DTMF encoded signal having a predefined
duration. In response to the CAS signal, the user's telephone system
enables a modem on the line to detect and receive the Caller ID
information. Because the CAS signal may be received during an ongoing
conversation, the DTMF detector attempting to identify the CAS signal must
have a high immunity to talk-off or talk-down errors.
As defined in proposed United States and international standards, the CAS
signal consists of the combination of a 2130 Hz tone and a 2750 Hz tone.
As defined by the proposed standard, the DTMF detector should accept the
CAS signal if the following parameters are met. The lower tone (2130 Hz)
and upper tone (2750 Hz) signals should be accurate in frequency within
.+-.0.5%. The CAS should have a dynamic range of -14 to -32 dBm per tone
and a power differential within the dynamic range of 0 to 6 dB between
tones. The duration of the CAS signal should be between 75 to 85 ms at the
user's equipment (i.e., customer premises equipment). At the central
office, the CAS signal tones should start and stop within 1 ms of each
other.
In general, a DTMF detector examines the line or communication channel for
the presence of two sinusoids using dedicated frequency domain algorithms,
including modified Goertzel algorithms, DFT/FFTs, auto-correlation, zero
crossing counting, and narrow band filter-based methods, among others.
A common technique for determining energy values in the frequency domain is
performed by transforming the time domain sampling data to frequency
domain data using a variant of the discrete Fourier transform (DFT) called
the Goertzel algorithm. The Goertzel algorithm is a second-order recursive
computation of the DFT using a feedback and a feed-forward phase. The
feedback phase computes a new output, y(n+1), for every new input sample
x(n), where N is the number of input samples. The feedback phase equation
is:
y(n+1)=c.multidot.y(n)-y(n-1)+x(n), Equation 1
where c is the Goertzel coefficient:
##EQU1##
In the Goertzel coefficient, .function. is the frequency to be detected and
F is the sampling frequency (i.e., 8 kHz).
The feed-forward phase is normally calculated after the N.sup.th data
sample is received (i.e., n=N). The feed-forward phase equation provides a
single output energy parameter, given by the equation:
.vertline.Yk(N).vertline..sup.2
=y(n).multidot.y(n)+y(n-1).multidot.y(n-1)-2.multidot.c.multidot.y(n).mult
idot.y(n-1), Equation 3
where
##EQU2##
and
##EQU3##
The values for N and k are integers. The choice of the values for k and N
for a given tone frequency requires a trade-off between accuracy and speed
of detection. As the number of samples, N, increases, the resolution in
the frequency domain increases, but the computation time for the
feed-forward phase increases. The spacing of the energy output values in
the frequency domain from the feed-forward phase is equal to half the
sampling frequency divided by N. If the tone being detected, .function.,
does not fall exactly on a selected frequency point (i.e., a frequency bin
defined by k and N), part of the energy associated with the frequency
point appears in the energy output value of an adjacent frequency point.
This phenomenon is referred to as leakage. To avoid leakage, the values of
k and N are chosen to attempt to center the tone being detected within a
frequency bin.
Once the DTMF detector has calculated the energy values for a given frame
of data, the DTMF detector performs various tests to ensure that the
detected energy meets the above listed criteria for the CAS signal.
Previous DTMF detectors have used various thresholding techniques to
determine the presence of valid DTMF tones. These prior techniques, to
various extents, lack the degree of noise immunity necessary to reliably
detect the presence of the CAS signal.
The present invention is directed to overcoming, or at least reducing the
effects of, one or more of the problems set forth above.
SUMMARY OF THE INVENTION
One aspect of the present invention is seen in a signal detector including
a transform unit and a signal alignment unit. The transform unit is
adapted to receive a plurality of frames of time domain input samples and
generate a plurality of energy values based on the frames of time domain
input samples. Each energy value is associated with one of the plurality
of frames. The signal alignment unit is adapted to identify at least a
first partial energy frame and a second partial energy frame from the
plurality of frames. The first partial energy frame is associated with a
first frequency and has a first energy value. The second partial energy
frame is associated with a second frequency and has a second energy value.
The signal alignment unit determines a misalignment between the first and
second partial energy frames based on the first and second energy values.
Another aspect of the present invention is seen in a method for detecting a
signal. The method includes generating a plurality of energy values based
on the frames of time domain input samples. Each energy value is
associated with one of the plurality of frames. At least a first partial
energy frame and a second partial energy frame are identified from the
plurality of frames. The first partial energy frame is associated with a
first frequency and has a first energy value. The second partial energy
frame is associated with a second frequency and has a second energy value.
A misalignment between the first and second partial energy frames is
determined based on the first and second energy values.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention may be best understood by reference to the following
description taken in conjunction with the accompanying drawings, in which
like reference numerals identify like elements, and in which:
FIG. 1 is a block diagram of a communications system in accordance with the
present invention;
FIG. 2 is physical block diagram of a dual-tone multifrequency detector of
FIG. 1;
FIG. 3 is a functional block diagram of a dual-tone multifrequency detector
of FIG. 1; and
FIG. 4 is a flow chart of a method for detecting signal misalignment in
accordance with the present invention.
While the invention is susceptible to various modifications and alternative
forms, specific embodiments thereof have been shown by way of example in
the drawings and are herein described in detail. It should be understood,
however, that the description herein of specific embodiments is not
intended to limit the invention to the particular forms disclosed, but on
the contrary, the intention is to cover all modifications, equivalents,
and alternatives falling within the spirit and scope of the invention as
defined by the appended claims.
DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS
Illustrative embodiments of the invention are described below. In the
interest of clarity, not all features of an actual implementation are
described in this specification. It will of course be appreciated that in
the development of any such actual embodiment, numerous
implementation-specific decisions must be made to achieve the developers'
specific goals, such as compliance with system-related and
business-related constraints, which will vary from one implementation to
another. Moreover, it will be appreciated that such a development effort
might be complex and time-consuming, but would nevertheless be a routine
undertaking for those of ordinary skill in the art having the benefit of
this disclosure.
Referring now to the Figures, and particularly, to FIG. 1, a block diagram
of a communications system 10 is provided. The communications system 10
includes a first customer installation 15, a central office 20, a second
customer installation 25, and a third customer installation 30. The first,
second, and third customer installations 15, 25, 30 are connected to the
central office 20 by subscriber lines 16, 26, and 31, respectively. The
first customer installation 15 includes a dual tone multifrequency (DTMF)
detector 35, a modem 40, and a display unit 42. The central office 20
includes a DTMF transmitter 45 and a modem 50.
For purposes of illustration, assume a first party at the first customer
installation 15 is communicating with a second party at the second
customer installation 25. The communication is routed through the central
office 20. The communication could be a voice call or a data call (e.g.,
computer, fax, etc.). During the communication between the first and
second parties, a third party at the third customer installation 30
attempts to place a call to the first customer installation 15. The
central office 20 detects the attempt made by the third party and sends a
customer alert signal (CAS) through the DTMF transmitter 45 over the
subscriber line 16. The central office 20 may also send an audible
subscriber alert signal (SAS) over the subscriber line 16 to audibly alert
the first party of the impending call. The SAS signal is typically a
single frequency tone (e.g., 440 Hz in the Unites States) that is
typically sent before the CAS signal.
The DTMF detector 35 receives the CAS signal, detects its presence, and
activates the modem 40 to monitor the subscriber line 16. Following the
CAS signal (i.e., after some delay), the modem 50 at the central office 20
transmits a digital data stream containing identifying information on the
third customer installation 30, presumably about the third party. The
identifying information may include the associated phone number, name of
the third party, date, time, or other relevant information. The modem 40
at the first customer installation 15 receives the digital data sent by
the modem 50 at the central office 20, and provides the identifying
information to the display unit 42. Based on the identifying information,
the first party may choose to place the second party on hold and accept
the call from the third party. Alternatively, the first party may choose
to ignore the call from the third party entirely, or call the third party
back at a more convenient time.
Referring to FIG. 2, a physical block diagram of the DTMF detector 35 is
provided. The DTMF detector 35 includes a receiver 75, a digital signal
processor 80, and a memory 85. The receiver 75 is coupled to the
subscriber line 16 to receive an analog signal and convert the analog
signal into digital format. The receiver 75 samples the analog signal and
utilizes pulse code modulation (PCM) or other suitable techniques to
produce corresponding digital data. The receiver 75 may include linear
analog to digital (A/D) converters (not shown).
It is noted that the received signal may be compressed or companded, and
thus digital data produced by the A/D converter (not shown) in the
receiver 75 may be companded (i.e., may comprise logarithmically
compressed digital data). As is well-known in the art, companding refers
to logarithmically compressing a signal at the source and expanding the
signal at the destination to obtain a high end-to-end dynamic range while
reducing dynamic range requirements within the communication channel. If
the signal is companded, the receiver 75 logarithmically expands the data
to a linear format. It is also contemplated that the DTMF detector 35 may
receive digital signals (i.e., linear or logarithmic) directly from the
subscriber line 16, and thus a receiver 75 is not required in the DTMF
detector 35 for analog to digital conversion.
A functional block diagram of the DTMF detector 35 is provided in FIG. 3.
The DTMF detector 35 includes the receiver 75 adapted to sample the signal
present on the subscriber line 16, a Goertzel Discrete Fourier Transform
(GDFT) unit 110 adapted to transform the time domain samples of the
receiver 75 to frequency domain energy values, a gain unit 115 adapted to
scale the output of the GDFT unit 110 based on any asymmetry in sample
sizes, a screening unit 120 adapted to perform a low-level thresholding
function, a cross-frame energy level unit 125 adapted to compare frequency
energy levels across frames of data, and a signal alignment unit 130
adapted to determine misalignment between the frames.
It is contemplated that any or all of the functions of the receiver 75,
GDFT unit 110, gain unit 115, screening unit 120, cross-frame energy level
unit 125, and signal alignment unit 130 may be accomplished by the digital
signal processor 80. The memory 85 may be a read-only memory, a random
access memory, or a combination thereof. The memory 85 may be employed to
store program instructions for execution by the digital signal processor
80. Also, the results of calculations completed by the digital signal
processor 80 may be stored in the memory 85. It is also contemplated that
a microprocessor (not shown) or other type of general purpose processor
(not shown) may be used to perform any or all of the functions of the
receiver 75, GDFT unit 110, gain unit 115, screening unit 120, cross-frame
energy level unit 125, and signal alignment unit 130.
In the illustrated embodiment, the receiver 75 samples at a frequency of 8
kHz, a common sampling frequency used in telephone systems. The time
domain sampling data provided by the receiver 75 is received by the GDFT
unit 110. The GDFT unit 110 determines energy output values
.vertline.Yk(n).vertline..sup.2 at various frequency bins as described
above.
Because the sampling frequency, F, is predetermined (e.g., 8 kHz), only the
number of samples N may be varied to center the tone frequency, .function.
within a frequency bin. The values of N used by the GDFT unit 110 in the
illustrated embodiment for each of the tones in the CAS signal and their
associated errors are shown below in Table 1.
TABLE 1
______________________________________
Absolute
Goertzel
f (tone)
N k (float)
k (integer)
Error in k
Coefficient
______________________________________
2130 169 44.99625 45 0.00375 -0.0204126297
2750 160 55.0 55
-1.111140466
______________________________________
Because the values of N are different for each of the tones in the CAS
signal, the energy output values (i.e., defined by Equation 6) should be
normalized prior to comparison. A gain adjust unit 115 receives the raw
energy output values from the GDFT unit 110 and scales the energy output
value for the 2750 Hz signal by the normalization ratio:
##EQU4##
The output of the gain adjust unit 115 is received by a screening unit 120.
The screening unit 120 evaluates the energy output values for each frame
(i.e., set of N samples) to determine if the CAS signal is present. The
DTMF detector 35 stores energy output values for six frames of data. The
current frame is referred to as frame 0. The DTMF detector 35 also stores
the previous frame (frame -1) and the four frames following the current
frame (frames 1 through 4).
The screening unit 120 performs a test to determine if the energy output
value in a frequency bin for the current frame (i.e., frame 0) passes a
predetermined screening threshold. The screening threshold is
intentionally set at a relatively low value (i.e., 10 in the illustrated
embodiment). The CAS signal may have been received by the DTMF detector 35
in the middle of the initial frame (frame 0). Accordingly, the energy
output values would only reflect a partial frame of CAS signal data, and a
valid CAS signal might be rejected if the screening threshold is set too
high. The screening unit 120 may pass a valid CAS signal, but it may also
pass an imitation signal, such as voice or noise. Further testing seeks to
reject the imitation signals in favor of actual CAS signals to improve the
noise immunity of the DTMF detector 35.
If the screening unit 120 detects a possible valid tone, the next two
energy output values (i.e., frames 1 and 2) for the frequency bin are
provided to the cross-frame energy level unit 125. Based on the timing of
the CAS signal, frames 1 and 2 should represent full energy frames of CAS
signal data, assuming a valid CAS signal is present. The cross-frame
energy level unit 125 evaluates the energy output values for frames 1 and
2 to determine cross-frame energy and twist measurements.
If both of the cross-frame energy level and twist tests pass, the DTMF
detector 35 proceeds with detection of the CAS signal. A method and
apparatus for performing the cross-frame energy level and twist tests is
described in U.S. patent application Ser. No. 09/159,143, entitled
"Cross-Frame Dual tone Multifrequency Detector," filed on the same date as
this application, and incorporated herein by reference in its entirety.
The signal alignment unit 130 evaluates the frames of data to determine the
amount of misalignment between the 2130 Hz and 2750 Hz signals. As stated
above, the energy in frame 0 typically represents a partial frame of data.
By comparing the partial energy in frame 0 for the 2130 Hz signal to the
partial energy in frame 0 for the 2750 Hz signal, an estimate of the
misalignment between the signals is determined. Based on the timing of the
CAS signal tones compared to the number of samples in a frame, the energy
values in frames 1 and 2 should represent full frames of data (i.e., N=169
samples for 2130 Hz and N=160 samples for 2750 Hz). Assuming the energy
detected represents a valid CAS signal, the energy present in the partial
frame should be from the CAS signal. The frame energy is proportional to
the square of the analysis length (i.e., the number of samples in the
frame). The equation for determining the number of samples in a partial
energy frame, based on this relationship, is:
##EQU5##
where E.sub.Partial represents the energy in the partial frame, and
E.sub.avg represents the average energy measured in full energy frames 1
and 2.
The number of samples in a partial energy frame attributable to the CAS
signal is compared between the two tones to detect misalignment
therebetween. The misalignment is calculated for the beginning and end
partial energy frames (i.e., frame -1 or 0 and frame 3 or 4).
Referring to FIG. 4, a flow chart of a method for performing the signal
misalignment twist measurements is provided. The signal misalignment test
is entered in block 200. The average full frame energy values for the 2130
Hz and 2750 Hz signals are calculated in block 205. The partial energy
frame numbers are identified in block 210. Depending on the time when the
CAS signal was received, it is possible that frame 0 may represent a full
energy frame. The average energy values are multiplied by a partial frame
scaling factor to determine a partial frame threshold. If the energy in
frame 0 is greater than the partial frame threshold, frame 0 is considered
a full energy frame. The energy in frame -1 is then evaluated to determine
the number of samples in its partial frame. The number of samples in frame
0 (N) is added to the partial number determined in frame -1 to generate
the sample count used for comparison with the partial energy frame for the
other tone. In the illustrated embodiment, the partial frame scaling
factor is about 0.5. Other partial frame scaling factors are contemplated,
depending on the specific application. A similar process is used to
determine which of frames 3 and 4 represent partial energy frames. It is
contemplated that the signal alignment unit 130 frames 0 and 3 may be
assumed to be partial energy frames and the partial frame scaling factor
may be omitted.
The partial frame sample counts (PFN) for the two tones are calculated for
the leading and trailing portions of the frame data in block 215. If the
partial energy frame is determined in block 220 to be frame 0 or frame 3,
no adjustment to the frame count is necessary. Otherwise, the full frame
sample count N for the tone is added to the partial frame sample count in
block 225 (i.e., PFN=frame 0 or 3 samples {N}+frame -1 or 4 samples
{PFN}). It is conceivable that the CAS signal tones may be misaligned
across frames (i.e., PF#=0 or 3 for one tone and -1 or 4 for the other
tone). The partial frame sample counts are determined for the start and
stop times of the CAS signal. The start and stop time sample counts are
determined for each tone.
Based on the sample counts, start time and stop time misalignments are
calculated in block 230 by subtracting the partial frame sample counts of
the two tones (i.e., MA=.vertline.PFN(2130)-PFN(2750).vertline.. The start
and stop time misalignments are compared to a threshold in block 235. If
either of the start and stop times exceeds the misalignment threshold, the
detection is abandoned in block 240. Otherwise, the detection is continued
in block 245. It is contemplated that additional tests may be performed
subsequent to the misalignment test.
Interference cause by imitation signals can affect the apparent starting
and or stopping times of the received CAS signal tones. Accordingly, the
misalignment threshold used for rejecting the detection based on frame
misalignment involves a tradeoff between immunity to imitation signals
(i.e., talk-off) and the ability to correctly detect the CAS signal in the
presence of imitation signals (i.e., talk-down). Also, the selection of
the partial frame scaling factor involves a similar trade-off.
In the illustrated embodiment, the misalignment threshold is about 70
samples (i.e., 8.75 ms). Other thresholds are contemplated, depending on
the specific application and the level of acceptable talk-down or talk-off
performance. For example, in an application where speech interference is
not expected, the misalignment threshold may be lowered to about 20
samples (i.e., 2.5 ms).
After completing the misalignment tests, the digital signal processor 80
may set a flag indicating a detection or may set a flag indicating no
detection. The flag may be an internal flag within the digital signal
processor 80 or a flag within the memory 85.
The operation of the signal alignment unit 130 in accordance with the
method of FIG. 4 is illustrated by the following examples. The examples
are provided for illustrative purposes and are not to be interpreted as
limiting the application of the present invention.
A first set of hypothetical energy values is illustrated below in Table 2.
The values of Table 2 were obtained by supplying a -32 dBm CAS signal to
the DTMF detector 35 having a duration of 80 ms.
TABLE 2
______________________________________
Frame # 2130 Hz energy
2750 Hz energy
______________________________________
0 30 30
1 37 36
2 35 36
3 27 32
______________________________________
The average energy values for frames 1 and 2 are 36 for both tones. For
ease of illustration, it is assumed that frames 0 and 3 are partial energy
frames (i.e., partial frame scaling factor=1). Note that because the
energy values are normalized by the gain unit 115, the sample value for
the 2130 Hz tone (i.e., N=169) is used as the number of samples in a full
frame for both tones.
The partial frame energies are determined using Equation 7. The partial
frame sample count for both tones in frame 0 is:
##EQU6##
Also, from equation 7, the partial frame sample counts for frame 3 are 146
(2130 Hz) and 159 (2750 Hz). The start times of the two tones are in
perfect alignment (i.e., both have 154 samples). The stop times are
misaligned by 159-146=13 samples, or 1.625 ms. The misalignment is less
than the 70 sample alignment threshold for the stop time, so the detection
continues.
A second set of hypothetical energy values is illustrated below in Table 3.
The values of Table 3 were obtained by supplying a -14 dBm CAS signal for
75 ms.
TABLE 3
______________________________________
Frame # 2130 Hz energy
2750 Hz energy
______________________________________
-1 82 36
0 2248
2170
1 2231
2204
2 2236
2256
3 330
372
______________________________________
The average energy values for frames 1 and 2 are 2234 (2130 Hz) and 2230
(2750 Hz). Using a partial frame scaling factor of 0.5, frame 0 is
determined to be a full energy frame for both tones. Also, frame 3 is
determined to be a partial energy frame for both tones. The misalignment
test is thus performed using frames -1 and 3.
The partial frame energies are determined using Equation 7. The partial
frame sample counts for the tones in frame -1 are 32+169=201 (2130 Hz) and
21+169=190 (2750 Hz). The partial frame sample counts for the tones in
frame 3 are 65 (2130 Hz) and 69 (2750 Hz).
The start times of the two tones are misaligned by 11 samples (i.e., 1.375
ms), and the stop times are misaligned by 4 samples (i.e., 0.5 ms). The
misalignments in the start and stop times are less than the 70 sample
alignment threshold, so the detection continues.
A third set of hypothetical energy values is illustrated below in Table 4.
The values of Table 4 were obtained by supplying a -9 dBm speech and music
signal to the DTMF detector 35.
TABLE 4
______________________________________
Frame # 2130 Hz energy
2750 Hz energy
______________________________________
-1 88 8
0 80 77
1 194
135
2 196
122
3 180
63
4 72 89
______________________________________
The average energy values for frames 1 and 2 are 195 (2130 Hz) and 128
(2750 Hz). Using a partial frame scaling factor of 0.5, frame 0 is
determined to be a partial energy frame for both tones. For the 2130 Hz
tone, frame 3 is determined to be a fill energy frame (i.e., frame 4 is a
partial energy frame). For the 2750 Hz tone frame 3 is determined to be a
partial energy frame. From the partial frame determination, it is evident
that the stop times are misaligned across frames. The misalignment test is
performed using frame 0 for the start time and frames 3 and 4 for the stop
time.
The partial frame energies are determined using Equation 7. The partial
frame sample counts for the tones in frame 0 are 108 (2130 Hz) and 131
(2750 Hz). The partial frame sample counts for the tones in frames 3 and 4
are 102+169=271 (2130 Hz) and 118 (2750 Hz).
The start times of the two tones are misaligned by 23 samples (i.e., 2.875
ms), and the stop times are misaligned by 153 samples (i.e., 19.125 ms).
The misalignments in the start times is less than the 70 sample alignment
threshold, but the misalignment in the stop times is greater than 70
samples, so the detection is abandoned.
The examples listed above illustrate the improved immunity of the DTMF
detector 35 to false detections produced by imitation signals, such as
voice and noise. Although the present invention is described in reference
to the detection of a specific DTMF signal, the CAS signal, the
application of the present invention is not so limited. The apparatus and
methods described herein may be applied to improve the immunity of other
in-band signaling detectors (not shown) to the presence of imitation
signals in tone detection applications. It is also contemplated that the
present invention may be applied to a multi-tone multi-frequency (MTMF)
detector for detecting any number of tones in a received signal (i.e., two
or more tones).
The particular embodiments disclosed above are illustrative only, as the
invention may be modified and practiced in different but equivalent
manners apparent to those skilled in the art having the benefit of the
teachings herein. Furthermore, no limitations are intended to the details
of construction or design herein shown, other than as described in the
claims below. It is therefore evident that the particular embodiments
disclosed above may be altered or modified and all such variations are
considered within the scope and spirit of the invention. Accordingly, the
protection sought herein is as set forth in the claims below.
Top